8 DEVELOPMENT OF SIUM CICUTAEFOLIUM. 



If the several categories into which these leaves have been divided 

 are examined critically, the following facts will be noted : 



1. They are not all of equal value. The second category (&) is only 

 a special case of the third (c). The fourth (d) is transitional between 

 (c} and (<?). The fifth (e) is composite, but not readily divisible. 



2. Some of the distinctions are qualitative, i. e., complexly quantita- 

 tive, as exemplified by (&), (c), (e), etc.; others are simply quantita- 

 tive, as, for example, (/), (g~), and (h~). 



3. They are not arranged in a logical linear series and are incapable 

 of being so arranged. They have been arranged in the order in which 

 they become dominant during development and without reference to 

 their logical relations. The undivided or irregularly cleft blade so 

 characteristic of the second nepionic leaf is the logical termination of 

 a line of development from the 3-parted leaf through the half 3-parted, 

 but the pinnate leaf with one pair of lateral leaflets, instead of joining 

 directly to this undivided form which it follows, also comes logically 

 from the 3-parted type. If the progress of the phylogeny of the species 

 is correctly indicated by this ontogenetic series it would appear that 

 after a gradual departure from the ancestral tripartite leaf to the undi- 

 vided leaf there had been a sudden jump or mutation to the opposite 

 extreme in which the clefts of the tripartite leaf became so exaggerated 

 as to leave the lateral lobes separated from the terminal by a space, thus 

 forming a pinnate leaf. It can not rightly be either affirmed or denied 

 that such mutation took place in the history of the species, but it appears 

 to me that any a,ttempt to interpret the details of ontogenetic develop- 

 ment as corresponding so definitely to phylogenetic development would 

 be pressing the analogy to a wholly unwarranted degree. 



In fig. i the variation in the several nepionic leaves from the first to 

 the eighth is represented graphically in the form of curves, the several 

 categories being treated as if of equal value. Although, as we have just 

 seen, this is not strictly true, the figure still allows the derivation of at 

 least one interesting conclusion there is a progressive lessening of the 

 variability from the first leaf onward. The modal class in the first nepi- 

 onic leaf contained only 28.2 per cent of the individuals, or, if we should 

 include in this class the special form of 3-parted leaf which has been 

 separated as a distinct category, it would still contain only 47.6 per cent 

 of the individuals, while the modal class in the eighth leaf included 

 nearly 73 per cent of the specimens, and the variability within the class 

 was very much less in the latter than in the former. 



There are several ways in which we may account for the great varia- 

 bility in the earliest leaf and the decreasing variation with subsequent 



